Osteological Evidence for Paraphyly of the Avian Order Caprimulgiformes (Nightjars and Allies) Gerald Mayr

Home , Caprimulgiformes, Macrodipteryx

J. Ornithol. 143, 82±97 (2002) ã Deutsche Ornithologen-Gesellschaft/Blackwell Wissenschaftsverlag, Berlin ISSN 0021-8375

Osteological evidence for paraphyly of the avian order Caprimulgiformes (nightjars and allies) Gerald Mayr

Forschungsinstitut Senckenberg, Sektion fuÈr Ornithologie, Senckenberganlage 25, D-60325 Frankfurt am Main, Germany; Email: [email protected]

Summary The first phylogenetic analysis of osteological characters in the Caprimulgiformes (nightjars and allies) provides strong evidence that this order is paraphyletic. Well supported is monophyly of a clade comprising Caprimulgidae, Nyctibiidae, Aegothelidae and the traditional Apodiformes (swifts and hummingbirds). Within this clade Nyctibiidae (potoos) and Caprimulgidae (nightjars) on the one hand, and Aegothelidae (owlet-nightjars) and the traditional Apodiformes on the other are sister taxa. Whereas monophyly of Nyctibiidae and Caprimulgidae is in accordance with most previous studies, monophyly of the so far poorlystudied Aegothelidae and the traditional Apodiformes has not been suggested by previous authors. The affinities of the Steatornithidae (oilbirds) and Podargidae (frogmouths) cannot be conclusivelyresolved at present.

Keywords: Phylogeny, Caprimulgiformes, Apodiformes, Aegothelidae, osteology.

Zusammenfassung Osteologische Hinweise auf Paraphylie der Vogelordnung Caprimulgiformes (SchwalmvoÈgel) Die erste phylogenetische Analyse der Caprimulgiformes (SchwalmvoÈgel) aufgrund osteologischer Merkmale liefert deutliche Hinweise auf eine Paraphylie dieser Ordnung. Gut gestuÈtzt ist die Monophylie eines Taxons, das Caprimulgidae, Nyctibiidae, Aegotheli- dae und die traditionelle Ordnung Apodiformes (Segler und Kolibris) beinhaltet. Inner- halb dieser Gruppe sind Nyctibiidae (TagschlaÈfer) und Caprimulgidae (Ziegenmelker) auf der einen Seite, sowie Aegothelidae (HoÈhlenschwalme) und die traditionellen Apodifor- mes auf der anderen Seite monophyletisch. WaÈhrend Monophylie von Nyctibiidae und Ca- primulgidae im Einklang mit den meisten bisherigen Untersuchungen steht, wurde eine Monophylie der bisher kaum untersuchten Aegothelidae und der traditionellen Apodifor- mes noch nicht vorgeschlagen. Die groûsystematische Stellung der Steatornithidae (Fett- schwalme) und Podargidae (Eulenschwalme) kann zur Zeit nicht uÈberzeugend begruÈndet werden.

Introduction 19th and early20th century,well before the concepts of modern phylogenetic systematics Virtuallyall of the currentlyrecognized avian were introduced (Hennig 1950), and some re- orders were established byanatomists in the main verypoorlydiagnosed. An especiallyill-

U.S. Copyright Clearance Center Code Statement: 0021-8375/2002/14301±0082 $ 15.00/0 G. Mayr ´ Paraphyly of Caprimulgiformes 83 defined order is the Caprimulgiformes which closelyrelated (see Sibley& Ahlquist 1990: traditionallyincludes five recent families: pp. 414±418 for a detailed review of the his- the holarctic Caprimulgidae (nightjars), the toryof classification). Cracraft (1981: 700) neotropic Nyctibiidae (potoos) and Steatorni- noted some morphological characters in order thidae (oilbirds), and the Australasian Aego- to support this traditional classification. How- thelidae (owlet-nightjars) and Podargidae ever, the characters he listed as synapomor- (frogmouths). All are crepuscular or nocturnal phies of Aegothelidae and Podargidae are ei- birds, and most feed on flying insects (the ther absent in stem-group Podargidae Steatornithidae are frugivorous). Caprimulgi- (ªdouble-notched sternum with long posterior form birds have a superficiallysimilar external processesº, see Mayr 1999, 2001), incorrect appearance which includes a wide beak with a (ªanterior iliac blades extremelywell devel- verylarge gape, and a soft plumage with cryp- opedº ± in the podargid genus Batrachostomus tic colouration. In their osteologyand internal the alae praeacetabulares ilii are verynarrow), anatomy, however, the members of this order or insufficientlydefined (ªpeculiar egg-white are remarkablydifferent. Monophylyof the protein patternº). In a recent paper, Cracraft Caprimulgiformes has never been convinc- (2001: Fig. 3) considered the systematic posi- inglyestablished with derived morphological tion of Aegothelidae and Podargidae to be un- characters, and the characters usuallylisted to resolved. diagnose this order (e. g. Stresemann 1927±34, Wetmore (1960) thought the Steatornithidae Sibley& Ahlquist 1990) are either not present to be the sister taxon of all other caprimulgi- in all taxa, plesiomorphic (e. g., anisodactyl form birds, which is reflected bytheir classifi- feet, holorhinal nares, presence of aftershaft), cation into a separate suborder. Cracraft (1981, or present in a wide range of other birds (e. g., 2001) assumed that this familyis the sister the pelvic muscle formula, arrangement of the taxon of Nyctibiidae and Caprimulgidae but, flexor tendons, nude oil gland). Despite this again, the characters he listed in order to sup- obvious lack of autapomorphic characters, the port monophyly of Steatornithidae, Nyctibii- traditional order Caprimulgiformes was con- dae, and Caprimulgidae are either not present sidered monophyletic by all recent and most in all three taxa (contraryto his statement, the earlier authors. Among the few exceptions are caudal margin of the sternum of the Nyctibii- Huxley(1867) who separated the Podargidae dae does not bear a single lateral process but from Nyctibiidae, Caprimulgidae, and Aego- exhibits two pairs of incisions) or plesiomor- thelidae, and Sharpe (1900) who separated phic (presence of processus basipterygoidei, Steatornithidae, Podargidae, and Aegothelidae see below). Curiously, Cracraft (1988: 354) from Nyctibiidae and Caprimulgidae in his lin- stated that ªthe steatornithids are seemingly ear classification. [. . .] primitive relative to the other four fami- The phylogenetic relationships between the liesº. caprimulgiform families also are poorlyre- Recent molecular phylogenetic studies do solved. Most authors based their studies on not support either monophyly of Caprimulgi- comparisons with Steatornithidae, Podargidae, dae and Nyctibiidae, or monophyly of Podargi- and Caprimulgidae, whereas published ac- dae and Aegothelidae. The DNA-DNA hybrid- counts of the anatomyof Nyctibiidaeand Ae- ization studies of Sibley& Ahlquist (1990) gothelidae are exceedinglyrare. Although only suggested monophyly of the taxon (Caprimul- Cracraft (1981) presented a formal phyloge- gidae + (Nyctibiidae + Steatornithidae)), with netic hypothesis based on morphological char- the Aegothelidae as the most basal taxon of re- acters, generallyNyctibiidaeand Caprimulgi- cent Caprimulgiformes. A phylogenetic analy- dae on the one hand, and Aegothelidae and sis using the mitochondrial cytochrome b gene Podargidae on the other were considered to be byMariaux & Braun (1996) showed the taxon 84 Journal fuÈr Ornithologie 143, 2002

(Nyctibiidae + (Caprimulgidae + Aegotheli- teros (2000) did confirm monophyly of Capri- dae)) to be monophyletic with the Steatornithi- mulgiformes and Apodiformes in most of the dae as the sister taxon of all other Caprimul- different trees that were presented (in the anal- giformes. Both studies supported an early yses of Bleiweiss et al. 1994, Mindell et al. branching of the Podargidae (which branch 1997, and Espinosa de los Monteros 2000 only just after the Aegothelidae respectivelyStea- the Caprimulgidae were included). tornithidae in the phylogenies of Sibley & Ahl- Further strong evidence for a close relation- quist and Mariaux & Braun). An analysis by ship between the apodiform and (at least some) Johansson et al. (2001) using nuclear DNA se- caprimulgiform birds comes from the well quences did not corroborate monophyly of known osteologyof the Aegialornithidae, an Steatornithidae, Podargidae, Nyctibiidae, and Eocene familyof basal swifts. The Aegialorni- Caprimulgidae (the Aegothelidae were not in- thidae (which are considered to be apodiform cluded in the analysis). However, the proposed byvirtuallyall recent authors, e. g. Olson phylogenetic positions of these taxa received 1985, Peters 1985, Karhu 1988, Mourer-Chau- onlyweak statistical support and are thus not vireÂ 1988a) not onlycloselyresemble the re- discussed in the following. cent Caprimulgidae and Aegothelidae in their Concerning their higher systematic position osteology(see Olson 1985 and below), but are most authors considered the Caprimulgiformes also verysimilar to the roughlycontemporane- to be most closelyrelated either to owls (Stri- ous extinct caprimulgiform familyArchaeo- giformes) or to swifts and hummingbirds trogonidae (Mourer-ChauvireÂ 1980, 1995). On (Apodiformes) (see Sibley& Ahlquist 1990). the other hand, there are no derived similarities However, morphological similarities between between earlyTertiaryowls and contempo- caprimulgiform birds and owls have mainly raneous caprimulgiform birds, nor derived been found if comparisons were made with Po- morphological characters which support dargidae and Steatornithidae (e. g., FuÈrbringer monophyly of owls and the traditional order 1888, Verheyen 1956), whereas the other fami- Caprimulgiformes. lies share derived characters with the Apodi- I attempted to analyze the exact phyloge- formes. Although the DNA-DNA hybridiza- netic relationships between apodiform and tion studies of Sibley& Ahlquist (1990) and caprimulgiform birds. The proposed relation- Bleiweiss et al. (1994) also resulted in mono- ships are shown in a cladogram, all nodes of phyly of the taxon (Caprimulgiformes + Strigi- which are supported with derived morphologi- formes), the methodologyemployedin these cal characters (Fig. 1). Due to the paucityof studies has repeatedlybeen criticized (e. g. published information which is available on Houde 1987, Lanyon 1992, Harshman 1994). the anatomyof Aegothelidae and Nyctibiidae, Moreover, the Fitch tree shown bySibley& the analysis is largely based on osteological Ahlquist (1990: Fig. 334) which according to characters Harshman (1994) ªshould be preferred to the Tapestry[= their more comprehensive trees in Fig. 353±385] in cases of conflictº did not Material and methods support monophyly of owls and caprimulgiform birds. A phylogenetic analysis using 12S Skeletons of the following caprimulgiform and apodiform taxa ± in addition to those of representatives mitochondrial rDNA byMindell et al. (1997) of all other higher avian taxa ± were examined in the did not support monophyly of either owls and collections of the Forschungsinstitut Senckenberg caprimulgiform birds or caprimulgiform and and the Museum fuÈr Naturkunde, Berlin: Steatorni- apodiform birds, whereas an analysis with thidae: Steatornis caripensis; Podargidae: Batra- complete cytochrome b and nearly complete chostomus septimus, Podargus strigoides; Aegothe- 12S rDNA sequences byEspinosa de los Mon- lidae: Aegotheles cristatus; Nyctibiidae: Nyctibius G. Mayr ´ Paraphyly of Caprimulgiformes 85

consistencyindex (CI), retention index (RI), and re- scaled consistencyindex (RC) were calculated. The robustness of the tree was tested with a bootstrap analysis. Given the nearly completely unresolved phylogenetic relationships between higher avian taxa, outgroup comparisons in the analysis with PA- UP are based on a hypothetical ancestor. Most of the characters used for the analysis could be clearly polarized into derived and primitive states; characters for which this was not possible have been coded as unknown in the outgroup.

Results Analysis with PAUP Analysis of the data set in Appendix 2 with PAUP 3.1 resulted in six most parsimonious trees (CI = 0.64, RI = 0.74, RC = 0.47). The topologyof the strict consensus tree is identi- Fig. 1. The proposed phylogeny of caprimulgi- cal to that shown in Fig. 1, except that the rela- form birds; the nodes are characterized in the text. tionships in the clade including Aegothelidae, Abb. 1. Die vorgestellten Verwandtschaftsbezie- Hemiprocnidae, Apodidae, and Trochilidae hungen caprimulgiformer VoÈgel; die Verzweigungs- are not resolved. Four trees exhibited the top- punkte sind im Text begruÈndet. ologyshown in Figure 1, in the two other trees the Trochilidae were the sister taxon to a clade griseus; Caprimulgidae: Caprimulgus europaeus, C. including Aegothelidae, Apodidae, and Hemi- pectoralis, C. carolinensis (onlyskull), Chordeiles procnidae. The 50 % majority-rule consensus minor, Macrodipteryx vexillarius; Apodidae: Apus tree of 200 bootstrapped replicates is shown in apus, Collocalia salangana, C. vanikorensis; Hemi- Fig. 2. procnidae: Hemiprocne comata; Trochilidae: Ama- zilia versicolor, Anthracothorax sp., Archilochus co- lubris, Chrysolampis mosqitus, Glaucis hirsuta, Monophyly of Caprimulgidae, Nyctibiidae, Melanotrochilus fuscus, Phaethornis pretrei, Tha- Aegothelidae, Hemiprocnidae, Apodidae, and lurania furcata, Trochilus polytmus. Details of the Trochilidae (Fig. 1, node 1). palatal region of Aeronautes saxatilis and Strepto- Since monophyly of the traditional Caprimul- procne zonaris (both Apodidae) were studied from the illustrations in Shufeldt (1885) and Cohn (1968). giformes has not hitherto been convincingly established with morphological characters Anatomical terminologyfollows Baumel & (Cracraft 1981, 1988), a sister group relation- Witmer (1993) and Vanden Berge & Zweers (1993). ship between the traditional Caprimulgiformes The phylogenetic hypothesis presented in this and Apodiformes is not supported, either. study(Fig. 1) was tested bya cladistic analysiswith Although Cracraft (1981: 700) stated that the phylogenetic software PAUP, version 3.1 (Swof- Caprimulgiformes and Apodiformes ªare de- ford 1993), using a data set of 25 anatomical charac- fined as forming a monophyletic lineage by ters (see Appendix 1 for descriptions of characters and Appendix 2 data matrix). The shortest tree was manyderived characters, particularlyin the found with the exhaustive search option and the skullº, all of the characters he listed as synapo- analysis was run with the delayed transformation morphies are absent in Steatornithidae and Po- (DELTRAN) mode, calculation with the ACCTRAN dargidae (ªveryshort, broad bill with the nasal mode did not change the resulting tree topology. The process of the premaxilla verythin and curved; 86 Journal fuÈr Ornithologie 143, 2002

ulum ventrale] to form ridge between themº in Steatornithidae and Podargidae ± I do not know what exactlyis meant bythe fourth character, ªlarge deep, rounded excavation under- cutting the humeral ridge and the ridge to the internal tuberosityº). Well supported bythe following characters, however, is monophyly of a taxon which includes Caprimulgidae, Nyctibiidae, Aegotheli- dae, Hemiprocnidae, Apodidae, and Trochili- dae (outgroup comparisons with palaeognathous and basal neognathous birds, e. g. Galliformes and Anseriformes, clearly suggest that these characters are derived within neognathous birds): (1) Ossa palatina with distinct processus rostrales (Fig. 3): These processes are absent in the Nyctibiidae in which the corresponding Fig. 2. 50 % majority-rule consensus tree of 200 area of the palatina is, however, stronglymodi- bootstrapped replicates based on a cladistic analysis fied (Fig. 3D). Distinct processus rostrales oth- of the data set in Appendix 2. The numbers indicate erwise onlyoccur in few other recent birds the percentage of bootstrapped replicates in which (see Huxley1867). the node was conserved. (2) Processus orbitalis of quadratum strongly Abb. 2. ¹50 % majority-ruleª Konsensus-Stamm- reduced (Fig. 4): This feature has been listed baum von 200 Bootstrap-Wiederholungen, basie- rend auf einer kladistischen Analyse des Datensat- as synapomorphic for the traditional Caprimul- zes in Anhang 2. Die Zahlen geben den Prozentsatz giformes and Apodiformes byCracraft (1981), der Bootstrap-Wiederholungen an, bei denen der but in Steatornithidae and Podargidae the proc- Verzweigungspunkt erhalten blieb. essus orbitalis is well developed. Contraryto the statement in Cracraft (1981), this process is however not completelylost in the Capri- the external nares verylarge and extending mulgidae (BuÈhler 1970: Fig. 5). With veryfew much the length of the bill; the jugal bars very exceptions (see Elzanowski et al. 2000), the long, thin, and joined to the premaxilla at the processus orbitalis is well developed in other far anterior end of the skull [. . .]; and the orbi- recent birds, including swallows (Passeri- tal process of the quadrate greatlyreducedº). formes, Hirundinidae), which have a similar Cracraft (1988) listed four other characters in bill shape to the above-listed taxa. order to support monophyly of the traditional (3) Condylus caudalis of quadratum reduced Caprimulgiformes and Apodiformes (without (Fig. 4): Concerning the Caprimulgidae, this mentioning those he alreadysuggested in feature was alreadynoted byBu Èhler (1970). 1981) but again, at least three of these are not The condylus caudalis otherwise is reduced in present in all taxa (the ªveryelongated and rel- onlyveryfew recent birds (see Elzanowski et ativelythinº articular surface of the condylus al. 2000). medialis of the quadrate and the ªdeep, very (4) Beak verywide with narial openings long, rectangular-shaped ligamental furrow on large and reaching almost to the tip of the humerusº are absent in the Podargidae, and the beak; distal part of rami mandibulae verynar- articular surface of the caput humeri is not row: owing to the predominantlynectarivorous ªcontinuous with internal tuberosity[= tuberc- diet of the Trochilidae, their beak is greatly G. Mayr ´ Paraphyly of Caprimulgiformes 87 elongated; unlike other nectarivorous birds with a similar bill shape (e. g. the passeriform Nectariniidae), however, the narial openings are verylong in hummingbirds, too. (5) Fossa dorsalis of the phalanx proximalis digiti majoris divided into two depressions by an oblique bulge: owing to the highlyderived morphologyof the corresponding phalanx, this feature cannot be clearlydiscerned in recent Trochilidae. In Caprimulgidae, Aegothelidae, and basal swifts (e. g. Mourer-ChauvireÂ 1988a, Karhu 1999) the phalanx is further bifenes-

Fig. 4. Right quadrate of Steatornis caripensis (Steatornithidae; A, B, C) and Aegotheles cristatus (Aegothelidae; D, E, F) in comparison. A, D: lateral view; B, E: caudal view; C, F: ventral view. The arrow indicates the pneumatic foramina on the caudal surface of the processus oticus in Aegotheles. Ab- breviations: p. ot. ± processus oticus; p. orb. ± processus orbitalis; c. caud. ± condylus caudalis; c. med. ± condylus medialis. Not to scale. Abb. 4. Rechtes Quadratum von Steatornis caripensis (Steatornithidae; A, B, C) und Aegotheles cristatus (Aegothelidae; D, E, F) im Vergleich. A, D: Lateralansicht; B, E: Caudalansicht; C, F: Ventralansicht. Der Pfeil zeigt auf die pneumati- schen Foramina auf der caudalen FlaÈche des Pro- cessus oticus in Aegotheles. AbkuÈrzungen: p. ot. ± Processus oticus; p. orb. ± Processus orbitalis; c. caud. ± Condylus caudalis; c. med. ± Condylus me- Fig. 3. Ventral view of caprimulgiform and apodi- dialis. Nicht im gleichen Maûstab. form skulls in comparison (slightlyschematic). A, Steatornis caripensis (Steatornithidae); B, Podargus strigoides (Podargidae); C, Chordeiles trated. Unquestionably, this, too, is a derived minor (Caprimulgidae); D, Nyctibius griseus (Nyc- feature which has apparentlybeen secondarily tibiidae); E, Aegotheles cristatus (Aegothelidae); lost in recent swifts (Olson 1985). In the Nycti- F, Apus apus (Apodidae). The large arrow indicates biidae the corresponding part of the phalanx is the processus rostralis, the small arrow points to the covered bya verythin osseous sheet, in the angulus caudolateralis; the asterisk marks the left Trochilidae there is a single fenestra. os palatinum. Not to scale. Abb. 3. Ventralansicht des SchaÈdels caprimulgi- In its composition, this clade is identical to former und apodiformer VoÈgel im Vergleich (leicht the ªCypselomorphaeº of Huxley (1867), who schematisch). A, Steatornis caripensis (Steatorni- was virtuallythe onlyauthor to make osteolog- thidae); B, Podargus strigoides (Podargidae); ical comparisons with both Aegothelidae and C, Chordeiles minor (Caprimulgidae); D, Nyctibius Nyctibiidae (though Huxley's study was re- griseus (Nyctibiidae); E, Aegotheles cristatus stricted to the palatal region). (Aegothelidae); F, Apus apus (Apodidae). Der gro- The earliest fossil record of the Podargidae ûe Pfeil markiert den Processus rostralis, der kleine Pfeil zeigt auf den Angulus caudolateralis; das is Masillapodargus from the Middle Eocene of Sternchen kennzeichnet das linke Os palatinum. Messel (Mayr 1999, 2001). This taxon shares Nicht im gleichen Maûstab. the characteristic large and massive bill with 88 Journal fuÈr Ornithologie 143, 2002

This is a unique and highlycharacteristic feature which is onlyfound in Nyctibiidaeand Caprimulgidae, and which to the best of my knowledge has not been mentioned in the liter- ature so far. Its functional significance is unknown. (2) Ossa palatina extremelywidened (Fig. 3C, D): In no other avian taxon are the ossa palatina widened to such a degree as in Nyctibiidae and Caprimulgidae. (3) Processus paroccipitales stronglyprotruding ventrally. (4) Mandible with intraramal joint and caudal half of rami mandibulae greatlywidened and dorso-ventrallyflattened (see BuÈhler 1970 for a detailed description of this feature in the Caprimulgidae). The combination of these fea- Fig. 5. Lateral view of skull of A, Chordeiles mi- tures is unique to Nyctibiidae and Caprimulgi- nor (Caprimulgidae) and B, Nyctibius griseus (Nyc- dae. tibiidae) in comparison (slightlyschematic). The (5) Caudal end of mandible unusuallysmall, arrow indicates the cone-like bonyprotrusion which with veryshort cotylalateralis and stout proc- is synapomorphic for Caprimulgidae and Nyctibii- essus medialis (see BuÈhler 1970: Fig. 15). This dae. Not to scale. Abb. 5. Lateralansicht des SchaÈdels von A, Chor- feature is unique to Nyctibiidae and Caprimul- deiles minor (Caprimulgidae) und B, Nyctibius gri- gidae. seus (Nyctibiidae) im Vergleich (leicht schema- Cracraft (1981) further listed a ªreduction in tisch). Der Pfeil zeigt auf den zapfenfoÈrmigen the length of the dentaryrelative to the post- knoÈchernen Vorsprung, der synapomorph fuÈr Capri- dentarybonesº as a synapomorphyofthe clade mulgidae und Nyctibiidae ist. Nicht im gleichen (Caprimulgidae + Nyctibiidae). Maûstab. Stem-group members of the Nyctibiidae are known from the Middle Eocene of Messel its recent relatives. An articulated skeleton (Mayr 2001). However, owing to the flattening from the earlyEocene of North America was of the bones, interpretation of the structure of described as a new taxon of the Steatornithidae the mandible in the fossil specimens is diffi- byOlson (1987), but the systematicposition of cult. Mayr (1999) assumed that the Eocene this bird is controversial (Mayr 2001, Mayr & forms had a less specialized mandible than re- Daniels in press). cent Nyctibiidae but a definitive assessment is onlypossible with better preserved specimens. Monophyly of Nyctibiidae Other earlyTertiaryNyctibiidaehave been and Caprimulgidae (Fig. 1, node 2) found in the Upper Eocene to Upper Oligocene deposits of the Quercy, France (Mourer-Chau- Osteological evidence stronglysupports vireÂ 1989), and the fossils reveal that the neo- monophyly of the clade (Nyctibiidae + Capri- tropic distribution of recent potoos is relictual. mulgidae). The following characters are un- The split between Nyctibiidae and Caprimul- questionablyderived and, apart from character gidae must have occurred before the early 3, unique among recent birds: Eocene, although the Paleogene record of the (1) Cone-like bonyprotrusion at the caudal Caprimulgidae still is veryscanty(Mourer- margin of the foramen nervi optici (Fig. 5): ChauvireÂ 1988b, Olson 1999). G. Mayr ´ Paraphyly of Caprimulgiformes 89

Monophyly of Aegothelidae, Hemiprocnidae, these characters are derived within neognathous Apodidae, and Trochilidae (Fig. 1, node 3) birds): With respect to their anatomy, the Aegotheli- (1) Os palatinum with greatlyprotruding an- dae are probablyone of the least studied avian gulus caudolateralis (Fig. 3, ªtranspalatine families which, owing to the scarcityof ad- processº of Lowe 1939): In the Trochilidae the equate material in museum collections, is usu- angulus caudolateralis is onlypoorlydevel- allyomitted from studies of caprimulgiform oped but bears a long cartilaginous extension relationships (e. g. Shufeldt 1885, Clark 1901, (Lowe 1939: pl. 4, Fig. 1A). A greatlyprotrud- Wetmore 1919, Lowe 1939, Verheyen 1956). ing angulus caudolateralis onlyotherwise oc- As far as I know, the bonypalate of Aegotheles curs in a veryfew recent birds (including Pass- is illustrated for the first time in the present eriformes, see Huxley1867). It is also present study(Fig. 3E). in the Podargidae, which, however, lack most other features characterising nodes 1 and 3 of In his pioneering studies of the avian palate, Figure 1. Huxley(1867: 469) alreadynoted that ª Ae- (2) Processus basipterygoidei reduced: Pro- gotheles approaches the Swifts more nearly cessus basipterygoidei are present in palaeog- than Caprimulgus does in the form of its pala- nathous and basal neognathous birds (e. g. Gal- tine bones, and in the absence of basipterygoid liformes, Anseriformes), and their absence is processesº. However, despite the striking simi- generallyconsidered to be derived within neo- larityin skull morphology(which was also gnathous birds (e. g. Ericson 1997). Although recognized byCohn 1968), so far no author these processes have been reduced independ- has considered the possibilitythat the Aego- entlyin a number of recent birds, among the thelidae are more closelyrelated to swifts than Caprimulgiformes theyare onlyreduced in to the other caprimulgiform birds. Obviously Aegothelidae and Podargidae (in the podargid this was largelydue to the fact that owlet- genus Batrachostomus vestigial remains of nightjars differ distinctlyfrom swifts in exter- these processes still are present). nal appearance and some features of the post- (3) Pneumatic foramina on the caudal sur- cranial skeleton, i. e. the proportions of wing face of the processus oticus (Fig. 4E): This and leg bones and the presence of deep inci- feature was listed byCracraft (1981) as a syna- sions in the sternum. However, since the corre- pomorphyof the taxon (Apodi + Trochili) but sponding features of apodiform birds (i. e. is also well developed in the Aegothelidae, of greatlyabbreviated humeri and ulnae, short which the entire quadratum is verysimilar to legs, and complete reduction of incisions in that of swifts and hummingbirds. In all other the caudal margin of the sternum) are unques- members of the traditional Caprimulgiformes, tionablyderived, their absence in the Aegothe- as well as in most other recent avian taxa (see lidae does not constitute evidence against Elzanowski et al. 2000), there are no pneu- monophyly of owlet-nightjars, swifts, and matic foramina on the caudal surface of the hummingbirds. processus oticus. I did not find anyderived characters shared (4) Coracoid, extremitas omalis hooked and byAegothelidae and the taxon (Nyctibiidae + processus lateralis greatlyreduced (Fig. 6): Caprimulgidae) that are not also present in the Unlike in all other caprimulgiform birds but Apodiformes. On the other hand, monophyly of like in all recent and fossil swifts, the extremi- Aegothelidae, Hemiprocnidae, Apodidae, and tas omalis of the coracoid is hooked in the Trochilidae is supported bythe following char- Aegothelidae, and the processus lateralis is acters (outgroup comparisons with palaeogna- greatlyreduced. This condition is unquestion- thous and basal neognathous birds, e. g. Galli- ablyderived within neognathous birds. The formes and Anseriformes, clearlysuggest that hooked extremitas omalis is absent in the Tro- 90 Journal fuÈr Ornithologie 143, 2002

of the extinct familyJungornithidae (Fig. 6B, C). (5) Cruciform origin of musculus splenius capitis: Burton (1971) first recognized that Ae- gothelidae, swifts and hummingbirds share a cruciform origin of musculus splenius capitis. This muscle arises from the neural spine of the second cervical vertebra, and in Aegothelidae, swifts, and hummingbirds ªthe fibres of the muscle take their origin on the opposite side of the midline from that on which theyinsertº (Burton 1971: 19). Although a tendencyto- Fig. 6. Right coracoid of A, Steatornis caripensis wards this feature is found in veryfew other (Steatornithidae); B, Aegotheles cristatus (Aegothe- avian taxa (though none of the other Caprimul- lidae); and the extinct apodiform taxa C, Jungornis giformes, see Burton 1971), the exaggerated tesselatus (Jungornithidae; after Karhu 1999) and D, Aegialornis gallicus (Aegialornithidae; after condition which is accompanied bya modified Mourer-ChauvireÂ 1988a) in comparison. The large neural spine of the second cervical vertebra is arrow points to the processus lateralis, the small ar- unique to the three above-mentioned taxa. row indicates the foramen nervi supracoracoidei. Judging from the shape of the neural spine of Not to scale. the second vertebra, this feature is absent in Abb. 6. Rechtes Coracoid von A, Steatornis cari- the Nyctibiidae (which is the only caprimulgi- pensis (Steatornithidae); B, Aegotheles cristatus (Aegothelidae); und den ausgestorbenen apodifor- form familythat was not investigated byBur- men Taxa C, Jungornis tesselatus (Jungornithidae; ton 1971). Cracraft (1981) listed this feature as nach Karhu 1999) und D, Aegialornis gallicus (Ae- synapomorphy of the Apodiformes. gialornithidae; nach Mourer-ChauvireÂ 1988a) im (6) Absence of caeca: like recent Hemiproc- Vergleich. Der groûe Pfeil zeigt auf den Processus nidae, Apodidae, and Trochilidae, Aegotheles lateralis, der kleine Pfeil markiert das Foramen ner- completelylacks caeca, which are verywell vi supracoracoidei. Nicht im gleichen Maûstab. developed in all other caprimulgiform birds (Beddard 1886, Lowe 1939). Caeca have been chilidae, but owing to their peculiar mode of reduced independentlyin a number of recent flight, hummingbirds exhibit a veryspecial- birds, but their absence in the Aegothelidae is ised extremitas omalis of the coracoid, in so much the more of phylogenetic significance which the processus acrocoracoideus is fused since Caprimulgidae, Nyctibiidae, Aegotheli- with the processus procoracoideus. The cora- dae, and swifts have a similar insectivorous di- coid of the Aegothelidae further corresponds et. with that of swifts and hummingbirds in that a In their postcranial osteology, the Aegotheli- foramen nervi supracoracoidei is present (see dae compare well to basal swifts of the extinct Fig. 6). This has not to myknowledge been familyAegialornithidae (e. g. sternum with noted byprevious authors ± FuÈrbringer (1888: slightlybifurcated spina externa, morphology 1342), for example, lists the absence of this of carpometacarpus, proximal end of the ulna, foramen as a diagnostic feature of the Capri- tibiotarsus, and hypotarsus, Fig. 7). Differen- mulgiformes. Although among ªhigher land- ces between Aegothelidae and Aegialornithi- birdsº (sensu Olson 1985) a foramen nervi dae mainlyconcern the much more abbrevi- supracoracoidei otherwise onlyoccurs in owls, ated humerus and ulna of the Aegialornithidae its presence might be plesiomorphic within re- (a derived character of swifts, see below) and cent birds. The coracoid of the Aegothelidae is the more elongated legs of recent Aegothelidae especiallysimilar to that of basal Apodiformes (a derived character of owlet-nightjars which G. Mayr ´ Paraphyly of Caprimulgiformes 91

Steatornithidae (i. e. vomer greatlyreduced or completelyabsent, and processus maxillopala- tini fused along their midline) but aegithognathous as in swifts (i. e. vomer well developed and with truncate rostral end), as alreadynoted byHuxley(1867). Although Lowe (1939: 330) considered the palate of the Trochilidae to be aegithognathous, too, the condition in hummingbirds is less clear; generallytheyare thought to be schizognathous (i. e. tip of vomer pointed). Depending on the author, the palate of Caprimulgidae and Nyctibiidae also is clas- Fig. 7. Wing bones of Aegotheles cristatus (Aego- sified either as aegithognathous or schizogna- thelidae; A, C, E) and Aegialornis gallicus (Aegia- thous (see BuÈhler 1970). lornithidae; B, D, F, after Collins 1976 and Mourer- ChauvireÂ 1988a) in comparison. A, B: right hume- Kitto & Wilson (1966) reported a unique rus; C, D: left ulna; E, F, right carpometacarpus and form of the enzyme malate dehydrogenase phalanx proximalis digiti majoris. All bones are which is shared byswifts and hummingbirds. shown in the same magnification. Unfortunately, these authors did not include Abb. 7. FluÈgelknochen von Aegotheles cristatus anycaprimulgiform taxon in their study,and it (Aegothelidae; A, C, E) und Aegialornis gallicus would certainlybe interesting to find out (Aegialornithidae; B, D, F, nach Collins 1976 und whether this feature is also present in the Ae- Mourer-ChauvireÂ 1988a) im Vergleich. A, B: rechter Humerus; C, D: linke Ulna; E, F, rechter Carpo- gothelidae. There is also verylittle published metacarpus und Phalanx proximalis digiti majoris. material on the myology and visceral anatomy Alle Knochen in derselben VergroÈûerung. of the Aegothelidae. Like apodiform birds, the Aegothelidae lack a propatagial slip of musculus biceps brachii which is present in the Cap- rimulgidae (this slip is, however, also absent in might be related to their fairlyterrestrial way Podargidae and Steatornithidae, see Beddard of living). Unlike that of the Aegothelidae, the humerus of the Aegialornithidae also exhibits a large processus supracondylaris, which is, however, onlypoorlydeveloped in other apodiform taxa (e. g. Hemiprocnidae and Jungor- nithidae). Although the sternum of the Aegothelidae is similar to that of the Hemiprocnidae in its proportions, it differs from that of all recent apodiform birds in the presence of four large fenes- Fig. 8. Sternum of A, Aegotheles cristatus (Aego- thelidae); B, Hemiprocne comata (Hemiprocnidae); trae in its caudal margin (Fig. 8). Whether the and C, Apus apus (Apodidae) in comparison. Note connection of the caudal ends of the trabeculae the additional and irregular perforations in the cor- byan osseous bridge is an earlystage of the pus sterni of Hemiprocnidae and Apodidae. Not to complete reduction of incisions found in swifts scale. and hummingbirds, cannot be said at present. Abb. 8. Sternum von A, Aegotheles cristatus (Ae- gothelidae); B, Hemiprocne comata (Hemiprocni- Contraryto previous statements (e. g. Ser- dae); und C, Apus apus (Apodidae) im Vergleich. venty1985, Sibley& Ahlquist 1990, Holyoak Beachte die zusaÈtzlichen unregelmaÈûigen Perfora- 1999), the palate of the Aegothelidae is not tionen im Corpus sterni von Hemiprocnidae und desmognathous as is that of Podargidae and Apodidae. Nicht im gleichen Maûstab. 92 Journal fuÈr Ornithologie 143, 2002

1886, Lowe 1939). Aegothelidae and apodi- Monophyly of Apodidae, Hemiprocnidae, and form birds further lack the musculus fibularis Trochilidae (Fig. 1, node 4) longus (this muscle is also absent in the Stea- Except for Cohn (1968), who proposed a clos- tornithidae but present in Caprimulgidae and er relationship to the Passeriformes, recent au- Podargidae, see Beddard 1886, Zusi & Bentz thors have considered hummingbirds to be the 1984). Apodiform birds are characterized bya closest extant relatives of swifts (e. g., Cracraft large number of additional derived non-osteo- 1981, 1988, Sibley& Ahlquist 1990). Mono- logical characters (e. g. FuÈrbringer 1888, Lowe phyly of swifts and hummingbirds is supported 1939, Karhu 1999) and future dissections of bya large number of derived myologicalfea- spirit specimens of the Aegothelidae, which tures (e. g., Lowe 1939, Zusi & Bentz 1984, are not available to me, will have to show Cracraft 1988, Karhu 1999), biochemical and whether some of these features are also present molecular analyses (Kitto & Wilson 1966, Sib- in owlet-nightjars. However, it has to be em- ley& Ahlquist 1990, Bleiweiss et al. 1994, Jo- phasized that the absence of derived apodiform hansson et al. 2001) and, among other features features in the Aegothelidae does not refute (see Cracraft 1988), bythe following derived the phylogeny proposed herein. Paraphyly of osteological characters: the taxon including Aegothelidae, Apodidae, Hemiprocnidae, and Trochilidae could onlybe (1) Proximo-dorsal part of narial openings shown byfinding derived characters which covered byan osseous sheet. more convincinglyassign the Aegothelidae to (2) Facies articularis coracoideus weakly some other avian taxon. saddle-shaped or convex (no sulcus as in other The traditional idea of a closer relationship birds), see Karhu (1999: 209). between Aegothelidae and Podargidae might (3) Incisions in the caudal margin of the ster- have been influenced bythe fact that the ru- num reduced. fous morphs of some Aegothelidae (e. g. Ae- gotheles insignis and A. tatei) have a similar (4) Humerus and ulna stronglyabbreviated plumage pattern to the rufous morphs of Po- and hand greatlyelongated. dargidae of the genus Batrachostomus. How- Further evidence for monophyly of the tradi- ever, no plumage feature is consistently tional Apodiformes comes from recentlyde- present in all members of Aegothelidae and scribed fossils from the earlyTertiaryof the Podargidae, and it seems more likelythat the Caucasus which show a mosaic of apodid and greyplumage, which is present in all species, trochilid characters (Karhu 1988, 1992, 1999). is primitive within the Aegothelidae. Although the Trochilidae lack some of the Although recent Aegothelidae and Podargi- features which diagnose nodes 1 and 3 of Fig- dae have a similar geographical distribution, ure 1, in each case the corresponding part of the fossil record shows that at least the Podar- the skeleton shows a highlyapomorphic mor- gidae were also present in the Eocene of Eu- phology(beak, extremitas omalis of coracoid, rope (Mourer-ChauvireÂ 1989, Mayr 1999, phalanx proximalis digiti majoris). 2001). I did not investigate the exact phylogenetic The earliest certain record of the Aegotheli- relationships between the traditional apodidae is an incomplete articulated skeleton from form taxa. Most authors have assumed mono- the Middle Miocene of Australia (Rich & phyly of Hemiprocnidae (tree swifts) and Apo- McEvey1977). However, since there is a fairly didae (true swifts), but Karhu (1988: 87, 1992: extensive fossil record of swifts from Eocene 383) considered a sister group relationship be- deposits, the Aegothelidae too must have tween Apodidae and Trochilidae (humming- branched off well before the earlyEocene. birds). G. Mayr ´ Paraphyly of Caprimulgiformes 93

Conclusion did not find derived characters which support monophyly of all taxa included in the analysis. Whereas paraphyly of the Caprimulgiformes The phylogenetic position of Steatornithidae has alreadybeen suggested bya small number and Podargidae can be settled onlybythe in- of earlier authors (see Sibley& Ahlquist clusion of a large number of other avian taxa, 1990), the phylogeny presented in this study which is beyond the scope of this study. As as- differs from all previous classifications in the sumed byprevious authors, these birds might proposed sister group relationship between be more closelyrelated to owls than to the cap- Aegothelidae and Apodiformes. Monophyly of rimulgid/aegothelid/apodid lineage. If this Aegothelidae and Apodiformes is strongly were the case, a nocturnal wayof living would supported by11 derived morphological char- have arisen onlytwice independentlyin Aego- acters. Most authors who have questioned a thelidae and Nyctibiidae/Caprimulgidae (if close relationship between caprimulgiform one does not accept nocturnal ancestors for the and apodiform birds have considered the latter Apodiformes). to be closelyrelated to the Passeriformes (perching birds) (e. g. Shufeldt 1885, Lowe 1939). However, the few derived osteological Acknowledgements characters which were listed byLowe (1939) I thank Andrzej Elzanowski (Universityof Wroclaw, Po- land), Stefan Peters (Forschungsinstitut Senckenberg), and in order to support monophyly of Apodiformes CeÂcile Mourer-ChauvireÂ (UniversiteÂ Claude Bernard, and Passeriformes are also present in the Ae- Lyon) for critical comments which improved the manu- gothelidae (e. g. aegithognathous palate, script. stronglyprotruding angulus caudolateralis [ªtranspalatine processº]), and Aegothelidae References and Apodiformes further correspond in some derived features in which the latter differ from Baumel, J. J. & Witmer, L. M. (1993): Osteologia. the Passeriformes (e. g. absence of caeca). In: Baumel, J. J., King, A. S., Breazile, J. E., Apart from the relative position of the Aego- Evans, H. E. & Vanden Berge, J. C. (Eds.): Hand- thelidae, the topologyof the cladogram in book of avian anatomy: Nomina Anatomica Avium. Publications of the Nuttall Ornithological Fig. 1 is identical to that presented byMariaux Club 23: 45±132. & Braun (1996) in their studyof the mitochon- Beddard, F. E. (1886): On the Syrinx and other drial cytochrome b gene (see introduction). Points in the Anatomyof the Caprimulgidae. Unfortunately, however, these authors did not Proc. Zool. Soc. Lond. 1886: 147±153. include the Apodiformes in their analysis. Bleiweiss, R., Kirsch, J. A. W. & LaPointe, F. J. Although the phylogeny of Sibley & Ahlquist (1994): DNA-DNA hybridization based phyloge- (1990) differs from that proposed in this study, nyfor ªhigherº nonpasserines: Reevaluating a in their figures 96, 106, and 107 the melting keyportion of the avian familytree. Mol. Phylo- genet. Evol. 3: 248±255. curve of Aegotheles is closer to that of the BuÈhler, P. (1970): SchaÈdelmorphologie und Kiefer- Apodiformes than are the melting curves of mechanik der Caprimulgidae (Aves). Zeitschrift the other caprimulgiform birds. fuÈr Morphologie der Tiere 66: 337±399. The phylogenetic relationships proposed in Burton, P.J. K. (1971): Some observations on the this studyseem to suggest that the Apodi- splenius capitis muscle of birds. Ibis 113: 19±28. formes evolved from nocturnal or crepuscular Clark, H. L. (1901): The pterylosis of Podargus: ancestors, since this assumption is more parsi- with notes on the pterylography of the Caprimul- gi. Auk 18: 167±171. monious than a four-fold independent origin of Cohn, J. M. W. (1968): The convergent flight mecha- nocturnal activity(in Steatornithidae, Podargi- nism of swifts (Apodi) and hummingbirds (Tro- dae, Nyctibiidae/Caprimulgidae, and Aegothe- chili) (Aves). PhD dissertation Universityof lidae). However, it has to be emphasized that I Michigan, Ann Arbor, Michigan. 94 Journal fuÈr Ornithologie 143, 2002

Collins, C. T. (1976): Two New Species of Aegialor- Karhu, A. (1988): Novoye semeystvo strizheobraz- nis from France, with Comments on the Ordinal nykh iz paleogena Yevropy [A new family of Affinities of the Aegialornithidae. In: Olson, S. L. swift-like birds from the Paleogene of Europe]. (Ed.): Collected Papers in Avian Paleontology Paleont. Jour. 3: 78±88. Honoring the 90th Birthdayof Alexander Wet- Karhu, A. (1992): Morphological Divergence within more. Smithson. Contrib. Paleobiol. 27: 121± the Order Apodiformes as Revealed bythe Struc- 127. ture of the Humerus. In: Campbell, K. E. (Ed.): Cracraft, J. (1981): Toward a phylogenetic classifi- Papers in Avian Paleontologyhonoring Pierce cation of the recent birds of the world (Class Brodkorb. Natur. Hist. Mus. Los Angeles Co., Aves). Auk 98: 681±714. Sci. Ser. 36: 379±384. Cracraft, J. (1988): The major clades of birds. In: Benton, M. J. (Ed.): The Phylogeny and Classifi- Karhu, A. (1999): A New Genus and Species of the cation of the Tetrapods, Volume 1: Amphibians, FamilyJungornithidae (Apodiformes) from the Reptiles, Birds: 339±361 Oxford. Late Eocene of the Northern Caucasus, with Cracraft, J. (2001): Avian evolution, Gondwana bio- Comments on the Ancestryof Hummingbirds. In: geographyand the Cretaceous-Tertiarymass ex- Olson, S. L. (ed.): Avian Paleontologyat the tinction event. Proc. R. Soc. Lond. (B) 268: 459± Close of the 20th Century: Proceedings of the 4th 469. International Meeting of the Societyof Avian Paleontologyand Evolution, Washington, D. C., Elzanowski, A., Paul, G. S. & Stidham, T. A. (2000): 4±7 June 1996. Smithson. Contr. Paleobiol. 89: An avian quadrate from the Late Cretaceous 207±216. Lance Formation of Wyoming. Journal of Verte- brate Paleontology20: 712±719. Kitto, G. B. & Wilson, A. C. (1966): Evolution of Ericson, P.G. P. (1997): Systematic relationships of malate dehydrogenase in birds. Science 153: the palaeogene familyPresbyornithidae(Aves: 1408±1410. Anseriformes). Zoological Journal of the Linnean Lanyon, S. M. (1992): Review of Sibley and Ahl- Society121: 429±483. quist's ªPhylogeny and Classification of Birds. Espinosa de los Monteros, A. (2000): Higher-level A Studyin Molecular Evolutionº. Condor 94: Phylogeny of Trogoniformes. Mol. Phylogenet. 304±310. Evol. 14: 20±34. FuÈrbringer, M. (1888): Untersuchungen zur Mor- Lowe, P.R. (1939): On the Systematic position of phologie und Systematik der VoÈgel, zugleich ein the Swifts (Suborder Cypseli) and Hummingbirds Beitrag zur Anatomie der StuÈtz- und Bewegungs- (Suborder Trochili), with special reference to organe, vol. 2. Amsterdam. their Relation to the Order Passeriformes. Trans- Harshman, J. (1994): Reweaving the tapestry: What actions of the Zoological Societyof London 24 can we learn from Sibleyand Ahlquist (1990)? (4): 307±348. Auk 111: 377±388. Mariaux, J. & Braun, M. J. (1996): A Molecular Phy- Hennig, W. (1950): GrundzuÈge einer Theorie der logenetic Surveyof the Nightjars and Allies Phylogenetischen Systematik. Berlin. (Caprimulgiformes) with Special Emphasis on Holyoak, D. T. (1999): Family Aegothelidae (Owlet- the Potoos (Nyctibiidae). Mol. Phylogenet. Evol. nightjars). In: del Hoyo, J., Elliott, A. & Sargatal, 6: 228±244. J. (Eds.): Handbook of the Birds of the World, Mayr, G. (1999): Caprimulgiform birds from the vol. 5: 252±265, Barcelona. Middle Eocene of Messel (Hessen, Germany). Houde, P. (1987): Critical evaluation of DNA hy- Journal of Vertebrate Paleontology19: 521±532. bridization studies in avian systematics. Auk 104: 17±32. Mayr, G. (2001): Comments on the osteology of Ma- Huxley, T. H. (1867): On the Classification of Birds; sillapodargus longipes Mayr 1999 and Paraprefi- and on the Taxonomic Value of the Modifications ca major Mayr 1999, caprimulgiform birds from of certain of the Cranial Bones observable in that the Middle Eocene of Messel (Hessen, Germany). Class. Proc. Zool. Soc. Lond. 1867: 415±472. N. Jb. Geol. PalaÈont. Mh. 2001: 65±76. Johansson, U. S., Parsons, T. J., Irestedt, M. & Eric- Mayr, G. & Daniels, M. (in press): A new short-leg- son, P.G. P. (2001): Clades within the `higher land ged landbird from the earlyEocene of Wyoming birds', evaluated bynuclear DNA sequences. J. and contemporaneous European sites. Acta Pa- Zool. Syst. Evol. Research 39: 37±51. laeontologica Polonica, 46 (3). G. Mayr ´ Paraphyly of Caprimulgiformes 95

Mindell, D. P., Dorenson, M. D., Huddleston, C. J., Peters, D. S. (1985): Ein neuer Segler aus der Grube Miranda, H. C. Jr., Knight, A., Sawchuk, S. J. Messel und seine Bedeutung fuÈr den Status der & Yuri, T. (1997): Phylogenetic Relationships Aegialornithidae (Aves: Apodiformes). Sencken- among and within Select Avian Orders Based bergiana lethaea 66: 143±164. on Mitochondrial DNA. In: Mindell, D. P. (Ed.): Rich, P.V. & McEvey, A. (1977): A new owlet- Avian Molecular Evolution and Systematics: nightjar from the earlyto mid-Miocene of eastern 213±247. Ann Arbor. New South Wales. Memoires of the National Mu- Mourer-ChauvireÂ, C. (1980): The Archaeotrogoni- seum of Victoria 38: 247±253. dae from the Eocene and Oligocene deposits of Serventy, D. L. (1985): Owlet-Frogmouth. In: ªPhosphorites du Quercyº, France. In: Campbell, Campbell, B. & Lack, E. (Eds.): A Dictionaryof K. E. (Ed.): Papers in avian paleontologyhonor- Birds: 421±422. Calton. ing Hildegarde Howard. Natural HistoryMuseum Sharpe, R. B. (1900): A hand-list of the genera and of Los Angeles County, Contributions to Science species of birds. Vol. 2. London. 330: 17±31. Shufeldt, R. W. (1885): Contribution to the Compa- Mourer-ChauvireÂ, C. (1988a): Les Aegialornithidae rative Osteologyof the Trochilidae, Caprimulgi- (Aves: Apodiformes) des Phosphorites du dae, and Cypselidae. Proc. Zool. Soc. Lond. Quercy. Comparaison avec la forme de Messel. 1885: 886±915. Cour. Forsch.-Inst. Senckenberg 107: 369±381. Sibley, C. G. & Ahlquist, J. E. (1990): Phylogeny Mourer-ChauvireÂ, C. (1988b): Le gisement du Bre- and classification of birds: A studyin molecular tou (Phosphorites du Quercy, Tarn-et-Garonne, evolution. New Haven and London. France) et sa faune de verteÂbreÂs de l'EoceÁne Stegmann, B. (1963): Der Processus internus indicis supeÂrieur. II Oiseaux. Palaeontographica (A) 205: im Skelett des VogelfluÈgels. J. Ornithol. 104: 29±50. 413±423. Stresemann, E. (1927±34): Aves. In: KuÈkenthal, W. Mourer-ChauvireÂ, C. (1989): Les Caprimulgiformes & Krumbach, T. (Eds.): Handbuch der Zoologie. et les Coraciiformes de l'EoceÁne et de l'Oligo- Berlin and Leipzig. ceÁne des Phosphorites du Quercyet description Swofford, D. L. (1993): PAUP: Phylogenetic analy- de deux genres nouveaux de Podargidae et Nycti- sis using parsimony, version 3.1. Illinois Natural biidae. In: Ouellet, H. (ed.): Acta XIX congressus HistorySurvey,Champaign. internationalis ornithologici: 2047±2055; Ottawa Vanden Berge, J. C. & Zweers, G. A. (1993): Myolo- (Universityof Ottawa Press). gia. In: Baumel, J. J., King, A. S., Breazile, J. E., Mourer-ChauvireÂ, C. (1995): Les Garouillas et les Evans, H. E. & Vanden Berge, J. C. (Eds.): Hand- sites contemporains (OligoceÁne, MP 25) des book of avian anatomy: Nomina Anatomica Phosphorites du Quercy(Lot, Tarn-et-Garonne, Avium. Publications of the Nuttall Ornithological France) et leurs faunes de VerteÂbreÂs. 3. Oiseaux. Club 23: 189±247. Palaeontographica (A) 236: 33±38. Verheyen, R. (1956): Les Striges, les Trogones et les Olson, S. L. (1985): The fossil record of birds. In: Caprimulgi dans la systeÂmatique moderne. Insti- Farner, D. S., King, J. R. & Parkes, K. C. (Eds.): tut royal des Sciences naturelles de Belgique 32 Avian Biology. 8: 79±238. New York. (3): 1±31. Olson, S. L. (1987): An earlyEocene oilbird from Wetmore, A. (1919): On the anatomyof Nyctibius the Green River Formation of Wyoming (Capri- with notes on allied birds. Proceedings of the mulgiformes: Steatornithidae). Documents des United States National Museum 54: 577±586. Laboratoires de GeÂologie de Lyon 99: 57±69. Wetmore, A. (1960): A classification for the birds of Olson, S. L. (1999): EarlyEocene Birds from East- the world. Smithsonian Miscellaneous Collec- ern North America: A Faunule from the Nanje- tions 139 (11): 1±37. moyFormation of Virginia. In: Weems, R. E. & Zusi, R. L. & Bentz, G. D. (1984): Myology of the Grimsley, G. J. (Eds.): Early Eocene Vertebrates Purple-throated Carib (Eulampis jugularis) and and Plants from the Fisher/Sullivan Site (Nanje- Other Hummingbirds (Aves: Trochilidae). Smith- moyFormation) Stafford County,Virginia. Vir- son. Contrib. Zool. 385: 1±70. ginia Division of Mineral Resources Publication 152: 123±132. Accepted: 30 May2001 96 Journal fuÈr Ornithologie 143, 2002

Appendix 1. Description of characters. 1. Skull, processus basipterygoidei: present (0), absent (1). Processus basipterygoidei are present in palaeognathous and basal neognathous birds (e. g. Galliformes, Anseriformes), and their occurrence is generallyconsidered to be plesiomorphic within neognathous birds (e. g. Ericson 1997). 2. Ossa palatina with greatlyprotruding angulus caudolateralis: absent (0), present (1). Outgroup comparisons with basal neognathous birds clearlysuggest that a greatlyprotruding angulus caudolateralis is derived (this feature onlyoccurs in a veryfew recent avian taxa). 3. Ossa palatina: not greatlyenlarged (0), greatlyenlarged (1). Greatlyenlarged ossa palatina (to the degree as shown in Fig. 3) are unquestionablyderived within neognathous birds and among their extant representatives onlyfound in Nyctibiidaeand Caprimulgidae. 4. Ossa palatina with long and slender processus rostrales: absent (0), present (1). Outgroup comparisons with basal neognathous birds suggest that long and slender processus rostrales are derived (this feature onlyoccurs in a veryfew recent avian taxa). 5. Cone-like bonyprotrusion at caudal margin of foramen nervi optici: (0) absent, (1) present. This is a unique feature among neognathous birds which is onlyfound in Nyctibiidaeand Caprimulgidae. 6. Beak verywide, with narial openings large and reaching almost to the tip of the beak: absent (0), present (1). Outgoup comparisons with palaeognathous and basal neognathous birds suggest that a verywide beak and greatlyelongated narial openings are derived. 7. Processus paroccipitales greatlyprotruding ventrally:absent (0), present (1). Outgroup comparisons with basal neognathous birds suggest that greatlyprotruding processus paroccipitales are derived. 8. Processus orbitalis of quadratum: not greatlyreduced (0), greatlyreduced (1). A greatlyreduced processus orbitalis is clearlyderived within neognathous birds (Elzanowski et al. 2000). 9. Condylus caudalis of quadratum reduced: absent (0), present (1). A reduced condylus caudalis is clearly derived within neognathous birds (Elzanowski et al. 2000). 10. Caudal surface of processus oticus of quadratum with pneumatic foramina: absent (0), present (1). This feature is veryrarelyfound in other neognathous birds and certainlyderived within that taxon (Elza- nowski et al. 2000). 11. Distal part of rami mandibulae verynarrow: absent (0), present (1). Outgroup comparisons with basal neognathous birds suggest that verynarrow rami mandibulae are derived. 12. Proximal end of mandible unusuallysmall, with veryshort cotylalateralis and stout processus medialis: absent (0), present (1). This feature is unique to Nyctibiidae and Caprimulgidae and certainly derived within neognathous birds. 13. Mandible with intraramal joint and caudal half of rami mandibulae greatlywidened and dorso-ventrally flattened: absent (0), present (1). The combination of these features is unique to Nyctibiidae and Capri- mulgidae. 14. Caudal margin of sternum: with four notches (0), with two notches or without notches (1). The presence of a four-notched sternum is generallyconsidered to be primitive within recent birds (e. g. Olson 1987). 15. Facies articularis coracoideus weaklysaddle-shaped or convex: absent (0), present (1). This feature is unique to swifts and hummingbirds; all other recent bird possess a sulcus articularis coracoideus (Karhu 1999). 16. Coracoid, extremitas omalis: not hooked (0), hooked (1). Outgroup comparisons with basal neognathous birds suggest that a hooked extremitas omalis is derived. 17. Coracoid, processus procoracoideus: (0) well developed, (1) greatlyreduced. The polarityof this feature is unknown. 18. Coracoid, foramen nervi supracoracoidei: present (0), absent (1). The polarityof this feature is unknown. 19. Coracoid, processus lateralis: not greatlyreduced (0), greatlyreduced (1). A greatlyreduced processus lateralis is certainlyderived within recent birds. 20. Humerus greatlyabbreviated (measuring less than 1.5 times the length of the coracoid), and ulna stout and short (measuring less than twice the length of the coracoid): absent (0), present (1). A greatlyabbreviated humerus and ulna are unquestionablyderived within recent birds, and otherwise onlyoccur in swallows (Hirundinidae, Passeriformes). 21. Fossa dorsalis of phalanx proximalis digiti majoris divided into two depressions byan oblique bulge: absent (0), present (1). Outgroup comparisons with basal neognathous birds suggest that the presence of this feature is derived. G. Mayr ´ Paraphyly of Caprimulgiformes 97

22. Phalanx proximalis digiti majoris, well-developed processus internus indicis (terminologyafter Steg- mann 1963): absent (0), present (1). A well developed processus internus indicis is absent in palaeognathous and basal neognathous birds and certainlyderived within recent birds (see Stegmann 1963). 23. Tarsometatarsus extremelystout and abbreviated (less than half as long as carpometacarpus): absent (0), present (1). An extremelystout and abbreviated tarsometatarsus (to the degree as found in Nyctibiidae and Steatornithidae) is unquestionablyderived within neognathous birds. 24. Caeca: well developed (0), absent (1). Outgroup comparison with other tetrapods suggests that the presence of caeca is plesiomorphic within recent birds. 25. Musculus splenius capitis: without cruciform origin (0), with cruciform origin (1). The cruciform origin is a derived feature which is onlyfound in a veryfew recent birds (see Burton 1971).

Appendix 2. Data matrix of 25 morphological characters for caprimulgiform and apodiform birds (see Ap- pendix 1. for definitions of characters). 0 = absent; 1 = present; ? = unknown states of characters. characteristics 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Ancestor 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ? ? 0 0 0 0 0 0 0 Aegothelidae 1 1 0 1 0 1 0 1 1 1 1 0 0 0 0 1 0 0 1 0 1 0 0 1 1 Nyctibiidae 0010111110111000110011100 Steatornithidae0000000000000100110001100 Caprimulgidae 0 0 1 1 1 1 1 1 1 0 1 1 1 1 0 0 1 1 0 0 1 1 0 0 0 Podargidae 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 Apodidae 1 1 0 1 0 1 0 1 1 1 1 0 0 1 1 1 1 0 1 1 1 1 0 1 1 Hemiprocnidae 1 1 0 1 0 1 0 1 1 1 1 0 0 1 1 1 1 0 1 1 1 1 0 1 1 Trochilidae 1 0 0 1 0 0 0 1 1 1 0 0 0 1 1 0 0 0 1 1 ? 1 0 1 1